On The Art of NDB DXing
by Sheldon Remington
©
1987-2000 All Rights Reserved
|
CHAPTER SEVEN: THE FREQUENCY DOMAIN
Part One - Modulation and Pitch The preceding chapters have dealt with the techniques required to achieve a good signal-to-noise ratio, so that we will hear literally hundreds of beacons filling the band when conditions are reasonable. At first, a DXer can simply continue to use his receiver the same way he has done on the broadcasting bands, with AM mode and wide selectivity. One quick discovery is that some beacons use voice weather forecasts for aviators. These are known as TWEB (Transcribed WEather Broadcast) outlets, and they play a continuously-repeating tape loop. Other beacons, called AWOS (Automated Weather Observation Stations) have a digitally derived voice giving a few particular details about atmospheric conditions at the beacon site or airfield it serves. Unfortunately we soon find out that nowhere on this tape does the beacon identify itself. This makes it of little interest for DXing, although it can be used for weather data if it is solid copy. Another complication in trying to identify TWEBS is that many of them are simply repeating another beacon's tape. In any case, the number of TWEB outlets is steadily declining, at least in the U.S., although the newer AWOS system is cropping up in increasing numbers. Existing ident-only beacons are now adding the digital-voice sound of AWOS, as it is quite cheap to maintain compared to TWEB or full-service weather information broadcasts on VHF. The other thing we hear from NDB's are tones of varying pitches--some of them steady, others going on and off. These beacons, of course, are identifying via slow-Morse telegraphy, and they are the means by which we identify and organize our DX catches. One way is to examine the tones in the time domain, which includes: translating the Morse into plain letters; noting whether there are long dashes or other special "formats"; timing the ident-cycle repetition rate, noting code speed; etc., and this time domain will be covered in detail in future chapters. The subject this time is the frequency domain, which is not sufficiently understood and appreciated by many veteran DXers, let alone newcomers. This is because the NDB band is so different from other bands in the type of modulation used by its transmitters, and therefore, in the equipment and technique needed to optimize our results. Just as with the time domain, there are several distinct sub-divisions to the frequency domain. Frequency Assignments First, let's look at the channelization used in assigning frequencies to beacons. Most countries simply use multiples of 1 kHz, including the U.S., Canada, and Australia. Actually, these countries mostly use 3 kHz channel spacing, but there are enough exceptions that 1 kHz spacing is what NDB DXers are dealing with. In many countries of Latin America, Oceania, Asia, and Africa, channels are normally in multiples of 5 kHz. For all the above channels, the frequencies listed in Ken Stryker's Beacon Guide and column can be taken as exact; incredibly enough, there appear to be no frequency errors in Ken's listings. A few countries, notably French territories, Japan, and the Russian Far East, put a small number of their beacons on "split" channels like 284.5. Ken Stryker rounds these frequencies down to the nearest kHz. Also, the European marine beacons are evidently utilizing a band-plan of channels 2.3 kHz apart, such as 287.3, 289.6, etc. Listings of these decimal frequencies can be found in official government NDB lists or in Jurgen Trochmcyk's Radio Beacon Handbook. All in all, though, 1 kHz (1000 Hertz) channelization is one of the basic facts of the NDB band. All of these frequency assignments are for the steady carrier of the beacon transmitter, rather than the Morse identifier pitch, which will be discussed later in this chapter. Although beacon transmitters are not held to the strict tolerances of broadcast transmitters, they are definitely crystal-controlled, as confirmed by many observations. Long-term carrier drift in non-defective beacons tends to be under 5 Hertz, and almost always within 20 Hertz of the assigned 1 kHz channel. The exceptions are in underdeveloped countries, and even those never seem to be found more than 200 Hertz from the assigned channel, although SJX in Michigan seems to be stretching things by occupying 381.5 kHz for years! Anyway, the bottom line is that the frequency shown in Ken's listings can be taken as basically exact, and even used as calibrators, within the tolerances above. Modulation Since the carrier signal normally just sits loudly and steadily at its point in the spectrum, you might rightly wonder who needs it? Well, not DXers, and our loggings would mushroom if the carriers were all removed from the band. The answer is that navigators with simple AM-mode receivers need to have a carrier. So, since the early days of radio, NDB's have used an obscure type of modulation known as tone-modulated AM telegraphy, abbreviated "A2." This produces the carrier signal, along with a pair of "sideband" signals which are those you hear with Morse Code. To illustrate, here is the frequency spectrum resulting when you modulate a carrier with a tone:
Note that there are two spots in the spectrum where there are identifier signals, quite distinct from the central carrier. These two signals do not necessarily have the same strength, depending on the tuning of the NDB's antenna system. Anyway, when an AM receiver is correctly centered on the carrier, the ident tones will be cleanly heard as a single pitch. But most of the transmitted power is consumed in the carrier, which is wasted on the listener with something other than an AM receiver. Worse, the radio spectrum is cluttered with signals on three frequencies, where one would suffice. In an attempt to reduce these drawbacks but still allow AM-mode reception, many beacons have adopted a modified modulation scheme. This consists of the carrier with just one sideband tone, as shown here:
In an AM receiver, this can still be tuned in and copied, albeit with somewhat more distortion. It does reduce the frequency clutter by 1/3, but it still wastes most of the transmitter power, from the standpoint of non-AM users or DXers. This scheme is used by most Canadian NDB's and by marine NDB's in the U.S. We'll return to A2 in a moment, but it's worth noting that a few countries, especially French overseas territories, have gone ahead and eliminated the carrier. Actually, what they've done is to merge the ident with the carrier, or simply key the carrier itself with the Morse code. This form of modulation is known as "A1," and it is the same CW mode in use on the ham bands and 1750 meters. It isn't audible on AM receivers, so users need to have a CW or SSB receiver (which generates its own substitute carrier). In this case alone, the ident frequency is identical to the "nominal carrier frequency" as shown in Ken's listings. Pitch Now, with A2 modulation, let's say we tune in using AM mode, and hear the ident tone of an NDB. Aside from its time-domain characteristics, it will have a certain fixed pitch, which I believe is the single most important parameter of beacons that we DXers can observe and share with each other. Even in AM mode, an unusual pitch results in a "clear channel" since differing pitches are easily isolated with filters (and distinguished by the ear/brain system). Remember that an A2 beacon occupies 2 or 3 points in the radio spectrum. The ident pitch is the same thing as the frequency difference between the carrier and each ident signal. If the ident pitch is 1000 Hertz (1 kHz), and the carrier is on 326 kHz, then the upper ident frequency would be on 327 kHz, and the lower ident frequency would be on 325 kHz. And if all idents had the same pitch, then we could always calculate the ident frequencies, since the carrier frequencies are accurately listed. But in the real world of DXing, ident pitches will be found covering at least the range from 250 to 1400 Hertz. And, lacking official data on beacon ident pitches, we've had to resort to measuring them ourselves. After several years of gathering such data, it is now possible to make a generalized table of NDB modulation pitches, at least for those countries within "earshot." Being a first compilation, it undoubtedly has some errors, and it is most certainly incomplete, so input is needed from DXers for future editions. In the table, note the preponderance of 1020 Hertz. This became a standard of sorts because early psychoacoustic experiments showed that around this pitch is where we have our greatest sensitivity to a tone in the presence of broadband random noise. Since this situation frequently applies in CW work on 1750 meters, DXers on that band might benefit from optimizing their receiving systems for an audio output pitch around 1000 Hertz. But, on the NDB band, the situation is much different, with close-in QRM being the limiting factor, so the strategy will have to be different there. Of course, in the AM mode we're forced to listen to the beacon at whatever ident pitch it's designed for. So any audio filter used in AM needs to be tunable over the 250-1400 Hertz range to cover the ident possibilities. But if we use the CW or SSB mode, we can decide what pitch we want. In the crowded NDB band, it is very helpful to use as low a pitch as we're comfortable with. This is because our ear/brain system hears frequencies logarithmically rather than linearly. To illustrate, say we have two idents, just 5 Hertz apart. At a pitch of 1000 Hertz, this 5 Hertz difference amounts to just 0.5%, which is too close to be discerned as different. So the two idents will sound like they are right on top of each other, and we'll have a hard time copying the weaker one. Now, if we reduce the pitch to 200 Hertz, then the 5 Hertz separation will amount to 2.5%, which is easily distinguished as 2 tones of differing pitch. And even if one ident is stronger, it won't drown out the other, since they'll be perceived as two different pitches. Thus, the lower our chosen pitch for listening, the more effective selectivity we will have; and selectivity is of utmost importance in NDB DXing. There are limits to how low our pitch can go, set mainly by the ear's reduced sensitivity at low frequencies, and by the difficulties in getting clean response from our audio reproduction systems. Depending on these factors, each DXer will have to experiment until he finds the best compromise pitch frequency for his ears and audio equipment. For example, Mitch Lee is using about 180 Hertz, even in non-QRM situations, because he feels that ear fatigue is reduced at low pitches. Mike Mideke currently listens at 400 Hertz, possibly because the headphones he uses have reduced output at lower frequencies. And this writer has settled on 240 hertz as his favored compromise. Using a fixed pitch for all NDB reception allows us to set up audio filters for a particular frequency. This can simplify the design if we are homebrewing our filters. It also reduces the number to tuning adjustments, of which we may already have several, in the antenna and RF stages. We are going to need some sort or reference source of audio pitch. When we adjust the pitch of the ident we are hearing (in CW/SSB mode) to match the reference, it is then centered in our system passband, and our calibration will be most accurate. Such a reference can be gotten in various ways. WWV and WWVH broadcast 500 and 500 tones in alternate minutes. Musical instruments are based on a standard pitch of 440 Hertz for the note "A," which means other notes will be simple fractions (220, 110, etc.) or multiples (880, 1760, etc.) of 440 Hertz. So you could use a household piano, organ, etc., or even buy an inexpensive homebrew oscillator such as a Radio Shack toy organ kit which can be kept at the receiver. Another way to get a fixed reference pitch is to tune (in CW/SSB mode) to the side of a reliably-heard carrier signal (NDB or otherwise) until it has the desired pitch, then enter it into one of the receiver's memories, assuming it is so equipped. INTERNATIONAL NDB IDENT PITCH LIST
No matter what pitch we settle on for listening, we should be sure our receiver audio systems are as clean as possible. The worst mistake is to DX with a speaker rather than headphones. Anyone who has run a slow sweep tone through a speaker indoors knows how extreme the deviations from a flat response are. Unless we put our ear right up to the speaker, we are listening to the composite response of the speaker and the room, which results in not only a rough frequency response, but also a smearing effect. This is worsened if we use speakers and rooms designed with radio in mind, because radio engineers are not usually competent in acoustical engineering. Even with headphones, the ones sold with radio listeners in mind are likely to have poor frequency and time response, not to mention comfort. It's worth spending, say, $50 or more to get a pair of high-fidelity phones. The sealed-ear types can be flat to below 20 Hertz, although there is a penalty in comfort. On the other hand, the lighter open-air types inevitably have a roll-off beginning somewhere above 100 Hertz, but it may be worth it if you can wear them comfortably all night; my personal recommendation would be Sennheiser HD424's, unless you DX in a noisy acoustical setting. Switching from a receiver's built-in loudspeaker to a set of audiophile headphones can give several valuable decibels of DX advantage. We have so far covered only a portion of the frequency-domain topics.
Next chapter will venture into the frequency "clustering" of idents, and
the special selectivity techniques needed to break up such clusters.
| |||||||||||||||||||||||||||||||||||||||||||||||||||
